I'm an AI/ML engineer mostly working around AI agents, agent harnesses, multi-agent systems, memory/context, and production AI workflows.
Right now I'm building:
- myagent.to — a messaging network that gives AI agents a permanent address so they can communicate across different owners and runtimes.
- OnWatch — a social listening tool that finds conversations about your product, competitors, and problem space, then filters them down to the ones worth acting on.
Previously, I worked on agent harnesses and AI co-workers for enterprise, AI hiring systems, candidate research agents, and AI SDR workflows.
In my free time, I like exploring new model architectures, diffusion-based language models, agent memory, and ways to reduce context bloat in long-running agents.
Forechain Supplier-risk intelligence powered by long-running agents that continuously monitor suppliers and surface evidence-backed risk signals.
Organic Hire AI-first ATS for small teams with resume parsing, candidate matching, ranking, and AI-assisted screening.
Open Session Open-source conference management software for CFPs, reviews, speaker onboarding, scheduling, and public agendas.
PagePod Multi-agent system for turning web content into podcasts.
Stock Insight Collaborative multi-agent system for stock research and market analysis.
- Built and owned an agent harness from scratch for Trugen's enterprise AI co-worker, powering products including ClaraSDR and Juno AI.
- Built AI-powered candidate search and hiring pipelines using embeddings, resume parsing, matching, and scoring at HireID.
- Built agents for deep research across GitHub and research profiles.
- Worked on Indian legal-domain LLMs and video generation research at AIR Center, VIT-AP.
- Worked across RAG, model fine-tuning, tool calling, multi-agent workflows, memory, context management, and production LLM systems.
- Multimodal Approach to Emotion Recognition using Deep Learning — ICIMIA 2023
- Paddy Crop Disease Detection using LeNet and MobileNet Models — INDIACom 2024
I'm always interested in talking about agent infrastructure, AI systems, model architecture, and weird agent ideas.